{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "
\n", "\n", "# Advanced Interval Plotting\n", "Author(s): Paul Miles | Date Created: July 18, 2019\n", "\n", "For the purpose of this example we will consider the Monod model demonstrated [here](Monod.ipynb)." ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "1.9.0\n" ] } ], "source": [ "import numpy as np\n", "import scipy.optimize\n", "from pymcmcstat.MCMC import MCMC\n", "import matplotlib.pyplot as plt\n", "import pymcmcstat\n", "print(pymcmcstat.__version__)" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "def model(q, data):\n", " x = data.xdata[0]\n", " a, b = q\n", " y = a*x/(b + x)\n", " return y.reshape(y.size,)\n", "\n", "def ssfun(q, data):\n", " yd = data.ydata[0]\n", " ym = model(q, data).reshape(yd.shape)\n", " return ((yd - ym)**2).sum()" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " Iteration Total nfev Cost Cost reduction Step norm Optimality \n", " 0 1 1.8012e-03 6.53e-02 \n", " 1 3 5.7019e-04 1.23e-03 2.18e+01 5.81e-03 \n", " 2 4 4.2786e-04 1.42e-04 3.68e+01 3.09e-03 \n", " 3 5 4.0841e-04 1.95e-05 7.61e+00 3.51e-04 \n", " 4 6 4.0839e-04 2.00e-08 7.33e-02 1.57e-07 \n", " 5 7 4.0839e-04 5.23e-12 4.69e-03 1.56e-10 \n", "`gtol` termination condition is satisfied.\n", "Function evaluations 7, initial cost 1.8012e-03, final cost 4.0839e-04, first-order optimality 1.56e-10.\n" ] } ], "source": [ "from pymcmcstat.MCMC import DataStructure\n", "data = DataStructure()\n", "# data structure\n", "x = np.array([28, 55, 83, 110, 138, 225, 375]) # (mg / L COD)\n", "y = np.array([0.053, 0.060, 0.112, 0.105, 0.099, 0.122, 0.125]) # (1 / h)\n", "data.add_data_set(x, y)\n", "# Calculate initial covariance matrix\n", "def residuals(q):\n", " yd = data.ydata[0]\n", " ym = model(q, data)\n", " res = yd - ym.reshape(yd.shape)\n", " return res.reshape(res.size, )\n", "\n", "ls0 = scipy.optimize.least_squares(residuals, [0.15, 100],\n", " verbose=2, max_nfev=100)\n", "theta0 = ls0.x\n", "n = data.n[0] # number of data points in model\n", "p = len(theta0) # number of model parameters (dof)\n", "ssmin = ssfun(theta0, data) # calculate the sum-of-squares error\n", "mse = ssmin/(n-p) # estimate for the error variance\n", "J = np.array([[x/(theta0[1]+x)], [-theta0[0]*x/(theta0[1]+x)**2]])\n", "J = J.transpose()\n", "J = J.reshape(n, p)\n", "tcov = np.linalg.inv(np.dot(J.transpose(), J))*mse" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n", "Sampling these parameters:\n", " name start [ min, max] N( mu, sigma^2)\n", "$\\mu_{max}$: 0.15 [ 0.00e+00, inf] N( 0.00e+00, inf)\n", " $K_x$: 49.05 [ 0.00e+00, inf] N( 0.00e+00, inf)\n", " [-----------------100%-----------------] 5000 of 5000 complete in 1.0 sec\n", "\n", "------------------------------\n", "name : mean std MC_err tau geweke\n", "$\\mu_{max}$: 0.1544 0.0220 0.0008 10.8093 0.9298\n", "$K_x$ : 61.7267 29.0816 1.2178 15.6448 0.7330\n", "------------------------------\n", "==============================\n", "Acceptance rate information\n", "---------------\n", "Results dictionary:\n", "Stage 1: 32.62%\n", "Stage 2: 49.36%\n", "Net : 81.98% -> 4099/5000\n", "---------------\n", "Chain provided:\n", "Net : 81.98% -> 4099/5000\n", "---------------\n", "Note, the net acceptance rate from the results dictionary\n", "may be different if you only provided a subset of the chain,\n", "e.g., removed the first part for burnin-in.\n", "------------------------------\n" ] } ], "source": [ "# Initialize MCMC object\n", "mcstat = MCMC()\n", "mcstat.data = data\n", "# Define model parameters, simulation options, and model settings.\n", "mcstat.parameters.add_model_parameter(\n", " name='$\\mu_{max}$',\n", " theta0=theta0[0],\n", " minimum=0)\n", "mcstat.parameters.add_model_parameter(\n", " name='$K_x$',\n", " theta0=theta0[1],\n", " minimum=0)\n", "mcstat.simulation_options.define_simulation_options(\n", " nsimu=int(5.0e3),\n", " updatesigma=True,\n", " qcov=tcov)\n", "mcstat.model_settings.define_model_settings(\n", " sos_function=ssfun,\n", " sigma2=0.01**2)\n", "# Run simulation\n", "mcstat.run_simulation()\n", "# Extract results and print statistics\n", "results = mcstat.simulation_results.results\n", "names = results['names']\n", "chain = results['chain']\n", "s2chain = results['s2chain']\n", "names = results['names'] # parameter names\n", "mcstat.chainstats(chain, results)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Plot Credible/Prediction Intervals\n", "Define function for generating intervals, setup calculations, and generate." ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\r", " [-----------------100%-----------------] 500 of 500 complete in 0.0 sec" ] } ], "source": [ "from pymcmcstat.propagation import calculate_intervals\n", "\n", "intervals = calculate_intervals(chain, results, data, model,\n", " s2chain=s2chain, nsample=500, waitbar=True)\n", "def format_plot():\n", " plt.xlabel('x (mg/L COD)', fontsize=20)\n", " plt.xticks(fontsize=20)\n", " plt.ylabel('y (1/h)', fontsize=20)\n", " plt.yticks(fontsize=20)\n", " plt.title('Predictive envelopes of the model', fontsize=20);" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Plotting\n", "Required inputs:\n", "- `intervals`: Output from `calculate_intervals`\n", "- `time`: Independent x-axis values\n", "\n", "Available inputs: (Defaults in Parantheses)\n", "- `ydata`: Observations, expect 1-D array if defined. (`None`)\n", "- `xdata`: Independent values corresponding to observations. This is required if the observations do not align with your times of generating the model response. (`None`)\n", "- `limits`: Quantile limits that correspond to percentage size of desired intervals. Note, this is the default limits, but specific limits can be defined using the ciset and piset dictionaries.\n", "- `adddata`: Flag to include data. (`False`, - if `ydata` is not `None`, then `True`)\n", "- `addmodel`: Flag to include median model response. (`True`)\n", "- `addlegend`: Flag to include legend. (`True`)\n", "- `addcredible`: Flag to include credible intervals. (`True`)\n", "- `addprediction`: Flag to include prediction intervals. (`True`)\n", "- `fig`: Handle of previously created figure object. (`None`)\n", "- `figsize`: (width, height) in inches. (`None`)\n", "- `legloc`: Legend location - matplotlib help for details. (`'upper left'`)\n", "- `ciset`: Settings for credible intervals. (`None` - see below)\n", "- `piset`: Settings for prediction intervals. (`None` - see below)\n", "- `return_settings`: Flag to return ciset and piset along with fig and ax. (`False`)\n", "- `model_display`: Model display settings. (See below)\n", "- `data_display`: Data display settings. (See below)\n", "- `interval_display`: Interval display settings. (See below)\n", "\n", "Default general display options:\n", "- `interval_display = {'linestyle': ':', 'linewidth': 1, 'alpha': 0.5, 'edgecolor': 'k'}`\n", "- `model_display = {'linestyle': '-', 'marker': '', 'color': 'r', 'linewidth': 2, 'markersize': 5, 'label': 'model', 'alpha': 1.0}`\n", "- `data_display = {'linestyle': '', 'marker': '.', 'color': 'b', 'linewidth': 1, 'markersize': 5, 'label': 'data', 'alpha': 1.0}`\n", "\n", "Display options specify to credible and prediction intervals:\n", "- `limits`: This should be a list of numbers between 0 and 100, e.g., limits=[50, 90] will result in 50% and 90% intervals.\n", "- `cmap`: The program is designed to “try” to choose colors that are visually distinct. The user can specify the colormap to choose from.\n", "- `colors`: The user can specify the color they would like for each interval in a list, e.g., [‘r’, ‘g’, ‘b’]. This list should have the same number of elements as limits or the code will revert back to its default behavior." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Case 1: Use default settings" ] }, { "cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "